allthose places. But the road from Hecatonpylos to Hyrcanial must decline to the north, for the city of Hyrcania lies somewhere between the latitudes of Smyrna and of the Hellespont.... Then the route runs on through Aria2 to Margiana Antiochia,3 first declining to the south (for Aria lies in the same latitude as the Caspian Gates), and then to the north, Antiochia being somewhere near the parallel of the Hellespont. Thence the road proceeds eastward to Bactra,4 and from that northward up the ascent of the hill country of the Comedi, and then inclining somewhat south through the hill country itself as far as the gorge in which the plains terminate. For the western end of the hill country is more to the north also, being (as Marinus puts it) under the latitude of Byzantium, the eastern end more to the south being under the latitude of Hellespont. Hence [the hills running _thus from south of east to north of west]. the road runs as he describes in the opposite direction, i.e. towards the east with an inclination south ; and then a distance of fifty scheeni extending to the Stone Tower would seem to tend northward. This Stone Tower stands in the way of those who ascend the gorge, and from it the mountains extend eastward to join the chain of Imaus which runs north to this from (the territory of) Palimbothra."...s
1 Joi n, N.W. of Astrabad.
2 The territory of Harah, Heri or HERAT.
:3 Supposed to be MARw. 4 BALKS.
5 I have not perhaps succeeded in rendering this description very intelligible. The old Latin versions and the Abb6 Halma's French translation seem simply to shirk the difficulties of the passage. I have not
access to any others or to Humboldt's Asie Centrale, which I believe con-
tains a dissertation on this route.
The account would perhaps be easier to understand if we knew more of the geography of the country towards Karategin, in which I suppose the hill country of the Comedi must lie. The chief difficulties arise in connexion with the expression as far as the gorge in which the plains terminate" (AWL T4s 4K5EX0µ4Vns Toi FeSla qiapâyyos), and the statement that fifty scheeni (one hundred and fifty miles ?) before reaching the Stone Tower the route lay northward. The former expression is intelligible if with Ritter we understand the passage of Imaus to have been that running from Kokand up the Jaxartes Valley to Andijan and across the Terek Daban to Kashgar, but in that case how could the route approaching the Stone Tower which he . places at Usx (where there are said to be ancient remains of importance) by any possibility run northward ? (see Ritter, vii, 483, 563; viii, 693). In the time of the Sui dynasty, or beginning of the seventh century, the Chinese knew three roads from Eastern into Western Turkestan, among which we naturally seek that of Maes Titianus. Of these three the first or north road seems from the description to have run north of the Thian Shan, and is out of the question ; a second or middle road passed from Kashgar to Farghana, and is no doubt that of the Terek Daban ; the third or south road passed through Khotan, and then through Chukiupo (said to be Yangihisar), and Kopantho (said to be Selekur or Sarikul ; see N. Ann. des Voy., 1846, iii, 47). Ritter takes the second for the route of Titianus, supposing the third route to be that by the Sirikul into Badakshan, which is certainly inconsistent with Ptolemy's data. But is. it certain that there was